Open Your Eyes and Ears: Leveraging Predictive Analytics to Improve Customer Experience

Beyond the Arc, Inc. © 2014 Beyond the Arc, Inc.

Customer Experience & Data Science

Open Your Eyes and Ears: Leveraging Predictive Analytics and Alternative Data

Sources to Improve the Customer Experience

Predictive Analytics World

Chicago, June 2014

Beyond the Arc, Inc.

…use data to assess your competitors’ strengths and weaknesses?

…detect and address regulatory risks before they result in costly fines?

…identify key customer pain points to improve the customer experience and drive retention?

Our analysis of the Consumer Financial Protection Bureau

(CFPB) complaint database, coupled with VOC insights

and social media analytics, helps identify key issues early

on and uncovers emerging trends to reduce risk and costs.

Imagine if you could…

2


Highlights and credentials

• Top 20 Most Promising Data Analytics Consulting Companies, CIO Review 2013

• Big Data 100, CIO Review, 2014

• Top Big Data Consultants, SourcingLine 2014

• Recognized by Forrester Research among 15 top firms in Customer Experience Consulting

• Featured speaker, Text Analytics World 2014

• Social media panelist, BAI Retail Delivery Annual Conference 2012

• Featured speaker on social media data mining, American Bankers Association 2011

Recent press

• Fox Business, January 2014

• BusinessWeek, April 2013

• Loyalty Magazine, Summer 2012

• Destination CRM, May 2012

• American Banker/Bank Technology News, Jan 2012

• BAI – Retail Banking Strategies, Jan 2012

• ABA Banking Journal, Jan 2012

• CUES, Credit Union Management, Jan 2012

• Social Business Today, Customer Experience Management @ TMC.net, and numerous blogs and online publications

AnalyticsStrategy Insights Action

3

A little about us


Here’s today’s agenda

• Big Data in the context of customer intelligence

• Creating customer insights: case studies

• Summary and key takeaways

4


• Regulatory reform is reshaping how you interact with your customer

• Consumer trust is largely a thing of the past

• Disruptive innovators, both established players and start-ups, covet your relationship with your customers

Winners will need to be agile and increasingly customer-centered

Nothing you don’t already know…your industry

is facing challenges

5


Inefficient Access1 in 2 don’t have access to the information across their organization needed to do their jobs

Lack of Insight1 in 3 managers frequently make critical decisions without the information they need

Inability to Predict3 in 4 business leaders say more predictive information would drive better decisions

Variety of Information

Volume of Digital Data

Velocity of Decision Making

Source: IBM Institute for Business Value

Organizations are operating with blind spots

6


Collect

Define

Report

Analyze

Improve

1. Define your objectives

2. Collect data sources and listen

3. Analyze, translating to business processes

4. Report on the findings

5. Improve based on insights revealed

7

Context is key to generating actionable insights


The business case for customer experience

transformation

• The shareholder value connection to customer experience is within reach, given that the value of a company is ultimately based on the amount of money customers are willing to spend with it.

--Peppers & Rogers Group 2012

• Customer experience leaders have more than a 16% advantage over laggards in consumers' willingness to buy more, their reluctance to switch business away, and their likelihood to recommend

--Tempkin Group

8


Success stems from understanding your

customers

• “Your most unhappy customers are your greatest source of learning.”

◦ Bill Gates

• “We aren’t in the coffee business serving people. We are in the people business serving coffee.”

◦ Howard Schultz, Starbucks

• “To understand the man, you must first walk a mile in his moccasin.”

◦ North American Indian proverb

• “This may seem simple, but you need to give customers what they want, not what you think they want.”

◦ John Ilhan (Australian entrepreneur)

9 9


Sometimes we just don’t feel our customers’

pain

10


InternalData from across the enterprise

o Customer feedback

o Frontline notes

o Transactional data

o Behavioral data

SocialSocial media about you and your competitors

o Twitter

o Facebook

o Blogs

RegulatoryCustomer complaints

o Federal

o State

o CFPB

11

Predicting customer issues requires a palette

of data sources and analytical approaches


Inbound calls to customer service

Customer service email

Blog posts

Letters to Federal/State regulators

Client verbatims

Survey responses

Employee surveys

Codified customer service data

Free form comments

Customer comments from advisors

UnsolicitedComments

SolicitedComments

Structured data Unstructured data

Fewer still are leveraging open data…

12

Most VOC programs struggle to incorporate

unstructured text, particularly in high volumes

Beyond the Arc, Inc.© 2013 Forrester Research, Inc. Reproduction Prohibited

• Surveys, relational database, XML, flat file, “internet of things”

• Data described by a schema

Structured text

• Free-form text

• Email, documents, tweets, blog comments, Facebook status, web forms

Unstructured text

• Audio, images, video

• Camera feeds, location data and maps, NoSQL/Hadoop

Hybrid

13

Examples

Big Data is about using all of your data


And employing methods that help you to

integrate it

Big Data

Online/Web

Social media

Operations

Transactions

Sensordata


Consumers are key in creating Big Data

• Over 34k likes are processed on Facebook each minute

• Over 27k new posts are captured on Tumblreach minute

• 48 hours of video is uploaded to YouTube every minute


Customer perspective • Customer surveys

• Online feedback forms

• Customer email

• Transaction & POS data

Frontline inputs• Feedback from customer-facing

teams

• Agent and customer service reps operating metrics

• Supervisors

Escalation queues• Internal escalations

• Team feedback

Do we have a customer focus?

Viewing the entire relationship?

Have we captured channel insights?

16

Look for data related to the customer journey,

across multiple channels and touchpoints






17


Case study: gaining insights from open data

• Bank accounts

• Bank services

• Credit cards

• Credit reporting

• Money transfers

• Mortgages

• Student loans

• Vehicle or consumer loans

CFPB has grown to include many products and services

18


A wide range of financial institutions now fall

within its purview

• Large banks and their affiliates

• Credit unions (above $10 billion)

• Nonbanks

• Residential mortgage companies

• Payday lenders

• Private education lenders

• Credit reporting agencies

19


Financial institutions have paid hefty fines

Institution Payment Date Reason

Capital One $210 million July 18, 2012 Deceptive marketing practices

Discover $214 million Sept. 24, 2012 Deceptive marketing practices

American Express $112.5 million Oct. 1 2012 Illegal credit card practices

JP Morgan Chase $389 million Sept. 19, 2013 Unfair billing practices

Ocwen $2.1 billion Dec. 19, 2013 Mortgage servicing abuses

20


Research shows the upside

Total revenue

potential

Churn reduced

Additional purchases

Word of mouth

Temkin $246M $91.8M $82.3M $72M

Forrester $307M $83M $213M $12M

• Incremental purchases from existing customers in the same year.

• Revenue saved by lower churn.

• New sales driven by word of mouth.

Increase in revenues for card issuers driven by improved experience

21


How does this affect banks, credit unions and

other financial institutions?

• Financial institutions are now subject to more stringent regulatory requirements

◦ Regulatory creep

• Consumers now have multiple avenues to voice complaints

• Companies need solid complaints management process to avoid costly regulatory exposure

• What if we could identify and address complaints before they become complaints, or even worse, regulatory issues?

22


Introduction to the CFPB Database

• In 2011, the CFPB made its database of consumer credit card complaints available to the public

• It was made public “to improve the transparency and efficiency of this essential consumer market” (CFPB)

• The CFPB database contains:◦ Almost 200,000 complaints

◦ Over 1,200 different companies

• http://www.consumerfinance.gov/complaintdatabase/

23

http://www.consumerfinance.gov/complaintdatabase/


Several key variables in the database can help

add depth to analysis

• Company

• Issue

• Submitted via

• Date received

• Zip code

• Company response

• Timely response?

• Consumer disputed?

24


We’ve completed an in-depth analysis of

consumer complaints

25View the full infographic at http://j.mp/1h0lEJE

http://j.mp/1h0lEJE


• Citi has the most complaints when scaled by cards in circulation

• Capital One starting to show signs of improvement

Source: CFPB, Beyond the Arc, Nilson report26

The database contains complaints about

almost 100 different credit card companies

2013


Source: CFPB, Beyond the Arc

Announcements of fines and regulatory action

lead consumers to “pile on” with complaints

27

Key takeaway: Managing your company’s reputation is more important than ever


0

2

4

6

8

10

12

January February March April May June July August September October November December

The 3 companies fined by the CFPB(in complaints per million cards)

Amex Capital One DiscoverSources: Nilson report 2012, CFPB, Beyond the Arc

Capital One and Amex experienced peaks in

complaints around the time of their fines

Paid $210M on 7/18

Paid $113M on 10/1

Paid $214M on 9/24

January 2012 – December 2012

28


Top 10 issues based on total database complaints

29

2013

Source: CFPB, Beyond the Arc,


The CFPB is likely most interested in complaints

that are closed with monetary relief

• These types of complaints are important because:

• The company has acknowledged responsibility, and

• Owes the customer financial restitution

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

Closed withexplanation

Closed withmonetary relief

Closed withoutrelief

Closed withrelief

Closed with non-monetary relief

Closed Untimelyresponse

In progress

Company responses as a percent of total

Source: CFPB, Beyond the Arc30


Billing disputes are the most prevalent issue with

monetary relief, but APR issues require attention

• The CFPB is likely paying attention to billing disputes since they account for the highest level of monetary relief

Source: CFPB, Beyond the Arc

0%

5%

10%

15%

20%

25%

30%

Billing disputes APR or interest rate Late fee Identity theft / Fraud /Embezzlement

Other fee

Top 5 issues that result in monetary relief

31


Beyond the Arc collects social media and

financial services data for analysis

• Our Social Customer Insights service:

◦ Contains over 350 million social media posts from– Twitter

– Facebook

– Industry blogs and forums

◦ Related to the top 100 banks and credit unions

◦ Over the past 24 months

◦ About 1,000,000 posts a day

32


0

10

20

30

40

50

60


Citibank - top 5 issues (number of complaints)

Billing disputes Interest rate Credit reporting Closing account Collection practicesSources: Nilson report 2012, CFPB, Beyond the Arc

We identified a correlation between Citibank APR

complaints and social media feedback


33


0

5

10

15

20

25

30

35

40


Citibank - Interest rate (number of complaints)

Interest rateSources: Nilson report 2012, CFPB, Beyond the Arc

Citibank APR complaints peaked in June


34


0%

2%

4%

6%

8%

10%

12%

14%

0%

5%

10%

15%

20%

25%

30%

35%

40%


Social Media and CFPB comparison (percent of total)

Social media CFPBSources: Nilson report 2012, CFPB, Beyond the Arc

Social media comments about Citibank APR show

a sharp spike in May


So

cia

l M

ed

iaC

FP

B

35

Social media spikes before complaints


• How's this for cynical capitalism. Rang up Citibank to pay-off and close my credit card. The lovely lady in the Phillipines asked me why and I said: your interest rate is far too High (21%) and she said Well I'm now willing to lower your rate because you are such a valuable customer .....Well if that's the case, why didn't you drop it earlier?‘

• Now that I am close to paying off almost all of my credit card debt, it feels so good to tell these credit card companies they have to drop my interest rate to something competitive… Citibank can kiss my a$$ with their terrible outsourced customer service.

Citibank APR sample comments

36


Capital One and Discover have greater proportion

of complaints about add-on products

• Together, the two companies were fined over $400 million

• Capital One and Discover account for a disproportionate number of complaints resolved with monetary relief for these products

0%

5%

10%

15%

20%

25%

30%

35%

Capital One Citibank Bank ofAmerica

JPMorganChase

GE CapitalRetail

Amex Discover Wells Fargo Barclays U.S. Bancorp

Percent of all complaints versus percent of marketingand add-on complaints resulting in monetary relief

% of all complaints

% of marketing and cardprotection servicescomplaints resulting inmonetary relief

Capital One was fined

$210 million on 7/18/12

Discover was fined

$214 million on 9/26/12

Source: CFPB, Beyond the Arc 37


Many companies have voiced their concern

about the CFPB and the complaint database

Concerns about the database:

• Complaints are not verified for accuracy

• Not representative of the population as whole

• Cannot determine if complaints are related to dissatisfaction or misunderstanding terms of service

• However, analysis of issues in the CFPB can help you determine where the customer experience is falling short

38


What is VOC?

An effective Voice of the Customer program uses text analytics to turn your unstructured data into structured data…

so you can quantify what your customers are talking about for analysis and take action

39


Analyzing customer data is important, but linking

insights to business processes generates ROI

Companies with advanced VOC

strategies

Companies withearly stage VOC strategies

The aim … is to know and understand the customer so well the product or service fits him and sells itself.

Peter Drucker

40


Voice of the Customer analysis is the key to

preventing feedback from escalating

• Voice of the Customer (VOC) analysis allows you to:

◦ Identify customer pain points to improve the customer experience and drive retention

◦ Detect regulatory risks andaddress them before the CFPB takes punitive action

◦ Tie VOC feedback to transactions,complaints, and other data sources,with Big Data and analytics

41


Of the top 10 trafficked sites in 2005 only 1 was

“social”

1 yahoo.com

2 msn.com

3 google.com

4 ebay.com

5 amazon.com

6 microsoft.com

7 myspace.com

8 google.co.uk

9 aol.com

10 go.com

1 google.com

2 facebook.com

3 yahoo.com

4 MSN/Bing

5 youtube.com

6 microsoft.com

7 aol.com

8 amazon.com

9 wikipedia.org

10 ask.com network

Source: Nielsen

Top Sites in 2005 Top Sites in 2013

42


of bloggers talk about brands on their blog

post brand or product reviews

70%

38%

Source: eConsultancy

People use social media to share product

experiences

43


• If Bank of America doesn't fix their Mint.com synch issue I am going to break out in hives. And switch banks. @BofA_Help@mint . (Twitter)

• @BofA_Help please get bofa working with mint.com again, i'dhate to have to switch to chase... . (Twitter)

• @BofA_Help ... To me the biggest plus of BofA is the mint.com integration. If that goes, you become a place to deposit checks. #replaceable . (Twitter)

• a Bank of America is considering capping debit card purchases at $50 or $100. Final reason to move my banking elsewhere. (Twitter)

Consumers often share very specific complaints

44


“Google never forgets” – Not even at the Greater

Alliance Credit Union in Hackensack, NJ

“I've been a member of this Credit Union for over 12 years, but tomorrow I'll ..close my accounts. Would have done it earlier, but I had a loan with them.”

“There's the hostile tone of voice over the phone from people who never tell you their name. There's also the belligerent attitude of most of the Hackensack folks and the sophomoric website that a high-schooler could have built.”

Posted on Yelp on 1/14/2009

Photo: HOWARD McWILLIAM/UK Telegraph

Found in Google search 11/4/2013

45

http://www.telegraph.co.uk/finance/personalfinance/consumertips/banking/6206351/Bank-complaints-dont-get-mad-get-even.html

http://www.telegraph.co.uk/finance/personalfinance/consumertips/banking/6206351/Bank-complaints-dont-get-mad-get-even.html






46


How do customer problems become

complaints?

47

Customer experiences a

problem

Customer provides feedback

Bank addresses problem

Resolution

Files complaint with CFPB

Closes account(s)

and / or Customers often try

multiple time to resolve their problems

Bank does not address problem


How can we identify and address these issues

so they don’t escalate?

48


problem



Resolution



Closes account(s)

and /or Customers often try



We need to focus on collecting, analyzing, and

tracking customer feedback

49


problem



Resolution



Closes account(s)

and /or Customers often try



Your workflow and processes will be critical to

success

1. Stage data in a shared analysis workspace

◦ Infrastructure: server, databases, network

◦ Tools: SQL, text analysis, data mining, statistical

◦ Collaboration: shared resources, common approach

2. Collect metadata

3. Characterize the data

◦ SQL tools: “usability”, descriptive statistics

4. Preliminary assessment of text mining suitability

5. Go/No-go decision for each data source

50

Cre

dit

: no

aa.g

ov

50


For data that makes the cut, you need a data

acquisition and management plan

6. Data load

◦ Initial pull of data

◦ Analysis environment vs. productionand reporting

7. Schedule periodic refresh

8. Automatic data exchange

◦ Data warehouse

◦ Sizing and IT resources

9. Data governance

◦ Who has access to data for analysis?

◦ Who can see output and reports?

51

Cre

dit

: bn

l.go

v

51


Our requirements for an analysis tool

• “Data flexibility”

◦ Multitude of formats: enterprise data warehouses, Excel, MS SQL Server databases, flat files, XML, and more

◦ Extensive data preparation– Each source required special handling

– “Create once, use forever”

• Integrated text analytics

• Statistical and linguistic models– Structured and unstructured data

• Quick ramp-up time for new users – Not rocket science

– Easy for us to train new users at the bank

52

Cre

dit

: nas

a.go

v

52


1 6 8

2 3 4 5

7

1) Identify Product discussed in

verbatim

2) Determine Detailed Category

using linguistic concepts

3) Discern Broad Theme via

linguistic concepts & statistical

analysis

4) Assess if Merger-related

5) Refine Detailed Category with

statistical analysis

6) Combine all findings using

business rules

7) Reporting module fuels

dashboards and reports

8) Deploy to Enterprise feedback

portals

Data Input

Our modeling uses both linguistic techniques and

statistical analyses to interpret and classify

53


What we liked in IBM SPSS text analytics

– Out of the box library resources, and the ability to build more

– Visual interface

– Variety of exploratory tools

– Ability to export as structured data, to use in predictivemodeling

Tools for categorization Explore links visually

Read sample comments on the fly

Review key concepts by volume and by type

54


Disclaimer and copyright

©2013, Beyond the Arc, Inc. All rights reserved

This document provides our commentary and analysis. The information contained herein is of a general nature and is not intended to address the circumstances of any particular individual or entity.

Although we endeavor to provide accurate and timely information, there can be no guarantee that such information is accurate as of the date it is received or that it will continue to be accurate in the future. No one should act on such information without appropriate professional advice after a thorough examination of the particular situation. No warranty expressed or implied.

The information provided in this presentation is not intended nor should be used as a substitute for legal advice or other expert opinions and services in specific situations.

This material may not be distributed.

Other company, product, and service names may be trademarks or service marks of others.

55


Thank you

Steven J. Ramirez, CEO


Office 1.877.676.3743

Email [email protected]

Digital beyondthearc.com

@beyondthearc

Facebook.com/beyondthearc

Slideshare.net/beyondthearc

56

Open Your Eyes and Ears: Leveraging Predictive Analytics to Improve Customer Experience

Data & Analytics

Transcript of Open Your Eyes and Ears: Leveraging Predictive Analytics to Improve Customer Experience